CA1295087C - Purification of recombinant beta-interferon incorporating rp-hplc - Google Patents

Abstract

Abstract of the Disclosure RP-HPLC methods for purifying recombinant beta-interferon are disclosed that employ wide-pore, silica-gel reverse-phase high performance liquid chromatograph (RP-HPLC) columns and solvent systems containing acetonitrile as the organic modifier and either heptafluorobutyric acid or trifluoroacetic acid as the organic acid.
The invention further concerns processes for purifying recombinant IFN-.beta. incorporating this RP-HPLC method.

Description

PURIFICATION OF RECOMBINANT BETA-INTERFERON INCORPORATING RP-HPLC

This invention is in the field of biochemical engineering.
More particularly, the inventlor concerns reverse-phase high performance liquid chromatography tRP-HPLC) methods for purifying recombinant beta-interferon (IFN-~), and improved processes for purifying recombinant IFN ~ incorporating said RP-HPLC methods. Said processes result in increased purity and decreased heterogeneity of 10 the product.
Naturally occurring interferons (IFNs~ are species-specific proteins, often glycoproteins, produced by various cells upon induction with viruses, double stranded RNAs, other polynucleotides, antigens and mitogens. Interferons exhibit multiple biological 15 activities such as antiviral, antiproliferative, immunomodulatory and anticellular funct10ns. At least three distinct types of human interferons have been identif~ed and characterlzed ~tl terms of thelr anti-viral, anti-growth and activation of natural killer cell (NK) activities. They are produced by leukocytes, lymphocytes, fibroblasts 20 and the im~une system and are classified as a, ~ and y interferons.
These are reported to be different proteins coded for by distinct structural genes.
Native human ~-interferon is generally produced by superinducing human fibroblast cultures with poly-IC (poly-25 riboinosinic acid and polyribocytidylic acid) and isolating andpurifying native human ~-interferon thus produced by chromatographic and electrophoretic techniques. Proteins or polypeptides which exhibit native ~-interferon-like properties may also be produced using recombinant DNA technology by extracting poly-A-rich 12S messenger RNA
30 from virally induced human cells, synthesizing double-stranded c-DNA
using ~he m-RNA as a template, introducing the c-DNA into an appropriate cloning vector, transforming suitable microorganisms with the vector, harvesting the bacteria and extracting the IFN-~therefrom. Nagola, S. et al., Nature, 284:316 (1980); Goeddel, D.~.
35 et al., Nature, 287:411 (1980); Yelverton, E. et al., Nuc. Ao;d Res., -s~8~7 9:731 (1981); Streuli, M. et al., Proc. Nat'l. Acad. Sci. (U~Sq),78:2848 (1981); European Patent Application Nu~bers 28033, published May 6, 1981; 321134, published July 15, 1981; 34307 published August 26, 1981; and Belgian Patent 837397, issued July 1, 1981 describe various currently used methods for the production of ~-interferon employing recombinant DNA techniques. The expressed proteins or polypeptides have been purified and tested and have been found to exhibit properties similar to those of native IFNs. Bacterially produced IFNs thus appear to have potential therapeutic use as 10 antiviral and anti-tumor agents and the production of IFNs by such bacterial ~ermentations is expected to yield sufficiently large quantities of IFN at a relatively lo~ cost of clinical testing.
Further, human IFN-~ genes have been altered by, for exa~ple, oligonucleotide-directed mutagenesis to produce IFN-~ protein 15 analogs thereof, such as the hu~an recombinant cysteine-depleted or cysteine-replaced interferon-~ analogs (muteins) disclosed in U.S.
Patent No. 4,588,585 issued May 13, 1986 to Mark et al. Specifically disclosed in that patent is the recombinant IFN-~ wherein the cysteine at position 17 is replaced by the neutral amino acld serine. That 20 IFN-~ analog is IFN~Bser17 Procedures for recovering and purifying bacterially produced IFNs are described in U.S. Patent Nos. 4,450,103; 4,315,852;
4,343,735; and 4,343,736; and Derynck et al., Nature (1980) 287:193-197 and Scandella and Kornberg, Biochemistr~, 10:4447 (1971).
25 Generally with these methods the IFN is not produced in a sufficiently pure form and in sufficiently large quantities for clinical and therapeutic purposes and the resulting IFN preparations produced by recombinant DNA techniques have residual amounts of chemicals and considerable microheterogeneity.
Purification and activity assurance of precipitated heterologous proteins is also described by U.S. Patent Nos. 4,5119502;
4,511,503; 4,512,922; and 4,518,526; and in European Patent 114,506.
European Publication No. 843,997 published March 257 1986 discloses a biochemical separation or recovery process in which 8~7 refractile bodies containing microbially produced IFN-~ are separated or recovered from the microorganism hosts, and further discloses protocols for then purifying the isolated refractile bodies or refractile material.
U.S. Patent No. 4,462,940 further discloses a process for purifying and formulati~g recombinant IFN-~.
U~S. Patent INo~ 43343,735 to Menge et al. teaches a process for the purification !f interferon by partitioning it in an aqueous multi-phase system in the presence of ion exchangers which are soluble in the system and are derivatives of polyethers.
U.S. Patent No. 4,289,690 to Pestka et al. discloses processes for purifying proteins including native human leukocyte interferon by utilizing one or more high performance liquid chromatography steps employing as solvents alkanols, such as n-propanol. See also Pestka et al., Pharmac. Ther., 29:299-319 (1985);
Langer et al., J. Investig. Dermatol., 83 (1):1285-1365 (1984); and Pestka, S., Archives Biochem. Biophys., 22I (1):1-37 (Feb. 15, 1983).
U.S. Patent No. 4,289,689 to Friesen et al. discloses how to recover and purify human native B-interferon by use of affinity chromatography and high pressure liquid chromatography.
In handling a biologically active protein, such as recombinant IFN-~, general considerations concerning the handling of proteins are relevant, including the necessity of preserving the protein's delicate tertiary structure in order to preserve biological activity, which requires the avoidance of denaturing pH conditions.
E. coli expressed recombinant IFN-~ and analogs thereof are insoluble in solutions which are at a pH range of 6 to 9. Therefore, various processes and additives have been devised to solubilize these proteins.
In producing a recombinant protein such as IFN-~ which is to be admi-nistered therapeutically to humans or animals, considerations of purity and homoge`neity of the final product are of the utmost concern. Reduction or elimination of minor IFN-~ species, and removal of both non-lFN-~ proteins and bacterial endotoxins, are of prime i~portance. Secondarily, in establishing or improvlng a purification scheme of a therapeutic, recombinant protein are considerations of efficiency and simplicity of the process.
The varia~ions on the theme of protein purification have been explored for more than fifty years. The literature on this subject is extensive and a plethora of techniques is available to the practitioner, including ion exchange chromatography, adsorption chromatography, gel electrophoresis, ammonium sulfate precipitations, and gel filtration. Over the years there have been substantial improvements in the technoloyy of conducting many of the foregoing methods, and in particular, it has been possible to auto~ate and speed up the procedures related to colu~n chromatography and develop~ent of electrophoresis gels. Despite these technical advances, and despite the large number of proteins which have been subjected to these procedures, the selection of a successful procedure, or more usually combination of procedures, for a particular protein found in a particular milieu has remained unpredictable, unselectable in advance, and subject to considerable experimentation in each particular case.
The instant invention incorporates a RP-HPLC method, specifically adapted for the purification of recombinant IFN-B, in purification processes essentialty outlined in European Publication No. ~439997 published March 25, 1986. References to purification schemes for recombinant proteins also employing RP-HPLC are listed 2S below.
U.S. Patent No. 4,485,017 issued to Tan et al. discloses a process for the isolation and purification of native interferons wherein a partially purified preparation of native interferon is sequentially passed through an antibody affinity column and a RP-HPLC
30 column.
U.S. Patent No. 4,569,790 issued to Ko~hs et al. discloses a process for recovering microbially produced IL-2 wherein the oxidized IL-2 is purified by RP-HPLC. A gradient solvent system co~prising an organic acid such as acetic acid or trifluoroacetic acid (TFA) and an .
.

~L~35 ~`~37 organic solvent such as propanol or acetonitrile is used to elute the IL-2 from the reverse-phase column.
Tarr et al., AnalO Biochem., 131: g9-107 (1983) describes the use of RP-HPLC resolution and recovery of cytochrome P-450 and bovine rhodopsin using ternary solvents, including acetonitrile and n-propanol and mixtures of the two.
Bennett et al., Biochem., 20:4530-4538 (1981) describes the purification of two major forms of rat corticotropin (ACTH) to apparent homogeneity by RP-HPLC using solvent systems containing either TFA or heptafluorobu~yric acid (HFBA) as hydrophobic counter-ions.
Bennett et al., Biochem., 197:391-400 (1981) describes the isolation and analysis of human parathyrin wherein RP-HPLC employing solvent systems composed of aqueous acetonitrile containing TFA or HFBA as hydrophobic ion-pairing reagents is used.
Smith-Johannsen et al., J. IFN. Res., 3 (4):473-477 (Mov. 4, 1983) discloses a procedure involving the chromatography of native human IFN-~ on an antibody affinity column followed by RP-HPLC. The RP-HPLC was performed wherein the elution solvent contained n-propanol in a linear gradient of between 0-100%, and wherein the C18 column was equilibrated with 100 ~M formic acid. See also Colby et al., J~
Immunol., 133 (6):3091-3095 (December, 1984) wherein recombinant IFN-~is purified according to the Smith-Johannsen et al. method.
Heukeshoven et al., Chromato~raphia, 19:95-100 (1985?
discloses RP-HPLC for various virus proteins and other hydrophobic proteins wherein a proportion of 60X formic acid in all solvents was used and wherein 2-propanol or acetonitrile was the organic modifier for gradient elution.
Burgess et al., PNAS (USA), 79:5753-5757 (October 1982) discloses the separation of two forms of murine epidPrmal growth factor by using RP-HPLC wherein the results of employing solvent systems containing either 0.2% trifluoroacetic acid or 0.2%
heptafluorobutyric acid and 20 or 50% acetonitrile were compared.

~L2~ 7 Friesen et al., Arch. Biochem. B~oph~ , 206 (2):432-450 (February 1981) dlscloses the pur~flcation of native human IFN-B by a comblnat~on of aff~nity chromatography and RP~HPLC whereln the reverse-phase column is equilibrated with pyridine-form~c acid containing isopropanol and/or n-butanol.
Molnar/(Ed.3, Practical Aspects of Modern HPLC (Proceedings Dec. 7-8, 1981!in West Berlin): Freisen, H., UHPLC of Protein on Reverse Phase xe~pl~f~ed w~th Human Interferons and Other Proteins:
Rev~ew and Scope of a MethodU (pp. 77-107~ provides a review article on reverse phase chromatography of proteins, including interferons.
The ins~ant invention provides improved purification processes ~ncorporating RP-HPLC methods specific for recombinant IFN-~. Said processes produce a hlghly pure IFN-~ product at very reduced levels of m~nor IFN-~ species, bacterial endotoxins and non-IFN-~
proteins. Sa~d processes prov~de flexible options for simplifylny andimproving the efficiency of the purificat~on of recombinant IFN-~, as well as increasing the purity and homogeneity of the product.
Accordingly, the present invention relates to d method for purifying recombinant interferon-~ (IFN-B) co~prising isolating the IFN-~ from hosts transformed to ~roduce it and passing the isolated IFN ~ through at least one~wide-pore, silica gel, reverse-phase hiyh performance liquid chromatography column using a solventsystem comprising an organic modi~ier comprising acetonitrile and an organic acid selected from the group consisting of heptafluorobutyric acid ~HFBA) and trifluoroacetic acid (TFA).
The inY2ntion further provides a process for recovering and purifying microbially produced IFN-~ comprising:
(d) disrupting the celi wall and cell membrane of a host m~croorganism cell culture transformed to produce said IFN-~;
(b) removing greater than 99% by weight of the salts from said disruptate;
(c) redisrupting the desalted uisruptate - 7 ~ ~29~7 (d) adding a material to the disruptate to increase the density or viscosity of, or to create a density or viscosity gradient in, the liquid witnin the disruptate;
(e) recovering refractile material containing the IFN-~ by high-speed centrifugation;
(f) solubilizing the IFN-~ in the refractile material with an aqueous solution of a solubilizing agent which forms a water-soluble complex with the recombinant IFN-~, said solution containing a reducing agent;
(g) oxidizing the product of step (f);
(h) passing the oxidized product of step (g) through a bonded, wide-pore, silica gel reverse-phase high performance liquid chromatography column using a solvent system comprising acetonitrile and either heptaEluorobutyric acid (HFBA) or triEluoroacetic ac:Ld (TFA);
(i) removing the solvent system from the product oE step (h), recovering the IFN-~ protein therefrom, and resolubilizing the IFN-~ protein in an appropriate buffer; and (j) further purifying the product of step (i).
Another aspect of this invention concerns recombinant IFN-~
purified by the RP-HPLC methods of this invention and IFN-~ recovered - and/or purified by the processes incorporating said RP-HPLC methods.
The bonded phase wide pore silica gel columns are preEerably alkane reverse phase columns, more preferably C~, C8 or C18 columns. When TFA is the acid used in the solvent system a C~ column is preferred.
Whereas, when HFBA is the acid chosen, a C18 column is preferred. The pore-size of said silica gels is preferably 300 angstroms or larger.
The elu~ion can be either isocratic or gradient, preferably gradient, either linear or non-linear, wherein the concentration ~v/v) of acetonitrile is in a range from 0% to 100%, preferably from 50% to about 80%, and more preferably from 50% to about 65%. The slope of the - 7a ~ 9~

gradient elution is preferably balanced with the amount of material loaded on the reverse phase column. The HFBA or TFA in the elution solvent system is in a concentration range of from about 0.001% to about 2%, preferably from about 0.05% to about 1%, and more preferably from about 0.1% to about 0.2%.
The most preferred RP-HPLC method of this invention comprises employing a C18 reverse-phase column with an elution solution containing acetronitrile as the organic modifier in a gradient oE from about 50%
to about 65%, and HFBA in a concentration of about 0.1%.
There are a number of processes for recovering recombinantly produced I~N-~. The RP-HPLC methods of the invention may be incorporated therein.
Preferred purification processes of the instant invention include those described in Eùropean Publication No. 8~3,997 published March 25, 1986 wherein the RP-HPLC methods of the instant invention are incorporated. Protocols 1 through 3, below, :Lllustrate preEerred processes whereLn the ~P-HPt,C methocls can be :Lncorporated. Such protocols schematica:Lly illustrate preferred processes Eor recovering, purifying and formulating microbially produced IFN-~.

,~

8 ~2~ &7 Fermentation Cel1 concentration Cell wall and membrane disruption homogenization 5 Diafiltration 5 ~M EDTA
Redisruption 2 mM EDTA; 1~ octanol (v/v);
homogenization Sucrose suspension 15-23% sucrose (w/w) Centrifugation 10,000 - 15,000 x g lO Paste solubilization 2% SDS, phosphate buffered saline Reduction 10 mM DTT; 2% SDS; 2mM EDTA;
pH 9; heat to 50C for 10 min.
under nitrogen; cool to about 25C; adjust pH to 7.4 with glacial acetic acid Organic extraction 2-butanol/suspension (v/v) Acid precipitation pH 6.2; 2 mM DTT; 0.1% SDS
Centrifugation 10,000 - 15,000 x 9 Acid precipitate 2% SDS; 5 mM EDTA; 50 mM
solubili~aiion phosphate buffer Reduction 20 mM DTT; pH 8.5; heat to 50 for 10 min. under nitrogen; cool to about 25C
Sephacryl~ S-200 column 50 mM acetate; pH 5.5; lg SDS;
1 mM EDTA
Oxidation - Iodosoben~oic acid tIBA) equimolar;
protein:IBA; Q.1~ SDSj 2 mM sodium -pyrophosphat ; pH 9; 1 mM EDTA
Concentration : pH 5.5 Sephacryl~ S-200 column 50 mM acetate; pH 5.5; 0.1~ SDS;
I mM EDTA

9 9L;~5~`~37 Concentrati on Sephadex0 G-75 column 50 mM acetate; pH 5.5; 0.1% SDS;
1 mM EDTA
Sephadex~ ¢-25 column 1 mM NaOH
Stabilizat.~on 1.25-5.00%, normal serum albumin ~ (human); pH 11 -~> 12.0 --> 7.5 Formulatipn 1.25% dextrose ~- Pre-filtration 0.45 ~M
Sterile filtration 0.22 ~M
Lyophilization Final Container Product Fermentat~oll Cell concentration 15 Cell wall and homogenization membrane di sruption Diafiltration- 5 mM EDTA
Redisruption 2 mM EDTA; 1% octanol (v/v);
homogenization 20 Sucrose suspension 15-23% sucrose (w/w) Centrifugation 10,000 - 15,000 x 9 Paste solubilization 1-2~ sodium laurate, 20 mM phosphate buffer, 50 mM DTT, pH 9-10 (with ` sonication~
25 Reduction 10 mM DTT; 1-2~ sodium laurate;
- : . 2 mM EDTA; pH 9; heat to 50C
~ for 10 min. under nitrogen; cool -: . to about 25C
: Sephacryl~ S-200 column 10 n~l Tris HCl; pH 9.2; 1-2%
: 30 sodium-laurate; 1 mM EDTA
' \
\, .:

Oxidation Iodosobenzoic acid (IBA) equimolar;
protein:IBA; 1-2~ sodium laurate;
2 mM sodium pyrophosphate; pH 9;
1 mM ED~A
5 Concentration pH 9.0 Sephacryl~ S-200 column 10 mM Tris HCl; pH 9c2;
0.1-0.5% sodium laurate; 1 mM EDTA
Concentration 10 mM Tris HCl; pH 9.2; 0.1 0.5%
sodium laurate, 1 mM EDlA
Sephadex0 G-75 column 10 mM Tris HCl; pH 9.2; ~1-.5%
sodium laurate; 1 mM ED1A
Sephadex~ G-25 column 0.1% laurate; 10 mM Tris HCl;
pH 9.2 pH Adjustment pH of eluate lowered quickly to 3 with 1.0 N HCl; sodium laurate precipitates Centrifugation and Filtration To remove precip~kated sodium laurate Stabilization 0.15% Trycol LAL-12 added Neutralization pH raised to between 6.0 and 7.2 20 Polyol addition 5X dextrose Pre-filtration 0.45 ~M
Sterile ~ ration 0.22 ~M
Lyophilization (immediate) Final Containèr Product Fermentation . . . ~ .
Cell concentration ~ ?
: Cell wall and . - homogenization membrane disruption 30 ~iafiltratioo 5 mM EDTA

11 ~295~37 Redisruption 2 mM EDTA; 1% octanol tv/v);
homogenization Sucrose suspension 15-23% sucrose ~w/w) Centrifugation 10,000 - 15,000 x 9 5 Paste solubilization 2% SDS phosphate buffered sali;ne Reduction 10 mM DTT; 2~ SDSj 2 mM EDTA;, pH 9; heat to 50C for 10 mi~.
under nitrogen; cool to abou 25C; adjust pH to 7O4 with glacial acetic acid Organic extraction 2-butanol/suspension (v/v) Acid preçipitation pH 6.2; 2 mM DTT; 0.1% SDS
Centrifugation 10,000 - 15,000 x 9 Acid precipitate solubilization 2% SDS; 5 mM EDTA; 50 mM
phosphate buffer Reduction 20 mM DTT; pH 8.5; heat to 50 for 10 min. under ni~rogen; cool to about 25C
Sephacryl~ S-200 column 50 mM acetate; pH 5.5; 1% SDS;
1 mM EDTA
Oxidation Iodosoben~oic acid (IBA) equimolari protein:IBA; 0.1g SDS;
2 mM sodium pyrophosphate; pH 9;
1 mM EDTA
.~5 Concentration pH 5.5 Sephacryl0 S-200 column 50 mM acetate; pH 5.5; 0.1% SDS;
1 mM EDTA
Concentration Sephadex~ G-75 column 50 mM acetate; pH 5.5; O.1X SDS;
1 ~M EDTA
Sephadex~ G-25 column . Ool~ sodium laurate (transfer component) in 10 mM:Tris-HCl, pH 9.2 :

12 ~ 7 pH Adjustment pH of eluate lowered quickly with 1.0 N HCl to pH 3; sodium 1aurate precipitates Centrifugation and To remove the precipitated sodium 5 Fi!tration laurate Stabilization 0.1~ Plurafac C-17 added ~éutralization pH raised ~o between 6.0 and 7.2 ~olyol addition 5X dextrose Pre-filtration 0.45 ~M
10 Sterile filtration 0.22 ~M
Lyophilization (immediate) Final Container Product Below, listed under the heading Alternative Downstream Purification Processes, are eight preferred downstream process alternatives incorporating the RP-HPLC methods of this invention.

Alternative Downstream Purification Processes Process I
1. Resolubilized Acid Precipitate (RSAP) 2. Gel filtration, preferably S-200 3. Contro11ed oxidation, preferably with IBA

4. RP-HPLC

5. Removal of organic solvents, preferably by precipitating the IFN-~protein, and recovery of the IFN-~, preferably, by resolubili7ing lt in an appropriate buffer 6. Gel filtration, preferably S-200 7. Formulation and optionally lyophilization 13 ~ 29S~87 Process II
The same as Process I except that the controlled oxidation step is performed after step 5 Process III
1. RSAP
2. RP-HPLC
3. Removal of organic solvents and recovery of IFN-~ protein 4. Controlled oxidation 5. Gel filtration, preferably S-200 6. Formulation and optionally lyophilization Process IV
The same as Process III except that the sontrolled oxidation step occurs prior to s~ep 2, that is, oxidation precedes RP-HPLC.
Process V
1. RSAP
lS 2. RP-HPLC
3. Removal of organic solvents and recovery of the IFN-~ protein 4. Controlled oxidation 5. Immunoaffinity purification step 6. Formulation and optionally lyophilization 20 Process VI
1. Solubilized particle paste from the abbreviated front-end process 2. Gel filtration, preferably by S-200 3. RP-HPLC
4. Removal of organic solvents and recovery of the IFN-~ protein.
30 5. Controlled oxidation 6. Gel filtration, preferably S-200 7. Formulation and optionaily lyophilization ~ 37 Process VII
1. RSAP
2. 6el filtration, preferably S-200 3. Controlled oxidation 4. Gel filtration, preferably S-200 5. RP-HPLC
6. Removal of organic solvents and recovery of ~he IFN-~ protein 7. Gel filtration, preferably G-75 8. Fnrmulation and optionally lyophili ation Process VIII
1. RSAP
2. ~el filtration, preferably S-200 3. Controlled oxidation 4. Gel filtration, preferably S-200 5. 6el filtration, preferably G-75 6. RP-HPLC
7. Removal of organic solvents and recovery of the IFN-~ protein 8. Formulation and optionally lyophilization For purposes of practicing the present invention, bacteria 20 are the preferred microorganism hosts, with E. coli being the most preferred.
In general, the recovery, purification and formulation - processes herein involve ~ermenting the host organisms transformed toexpress the I~FN-~, disrupting the cell wall and cell membrane of the 25 host organism, separating the refractile material containing the recombinant IFN-~ from the rest of the cellular debris, solubilizing the refractile material in an aqueous buffer under reducing conditions9 extracting the IFN-~ with 2-butanol or 2-methyl-?-butanol, subjecting the extracted IFN-~ to chromatographic purification, and - 15 - ~ %95~7 then diafiltering or desalting, preferably desalting, the IFN-~ to remo-ve the solubilizing agent, optional]y using a suitable transfer component, and formulating the purified IFN-~ with human serum albumin with or without dextrose or human plasma protein fraction (PPF) or with non-ionic biodegradable polymeric detergents.
Protocols 1 through 3 illustrate the processes in which the RP-~LC methods can be incorporated from culturing microorganisms transformed to produce IFN-~ protein in an appropriate fermentation medium through the final steps wherein the purified IFN-3 protein is stabilized and may then be lyophilized into therapeutic formulations that can then be reconstituted at the clinician's option. A preferred process Eor recovering and purifying recombinant IFN-~ comprises:
(a) recovering a refractile material containing microbially produced IFN-~ from a host m:lcroorganism cell culture transfolmed to produce said prote:Ln by the steps oE:
:L. disrupt:Lng the ce:ll wa:l.:L and cell membrane o.E the microorganism;
2. removing greater than 99% by weight oE the salts from said disruptate;
3. redisrupting the desalted disruptate;
4. adding a material to the disruptate to increase the density or viscosity of, or to create a density or viscosity gradient in, the liquid within the disruptate; and 5. separating the refractile material from the cellular debris by high-speed centrifugation;
(b) solubilizing the recombinant IFN-~ in the refractile material with an aqueous solution of a solubilizing agent which forms a water-soluble complex with the recombinant IFN-~, said solution containing a reducing agent;
(c) separating the recombinant IFN-~ from the resulting solution in the presence of said reducing agent by gel filtration;
(d) oxidizing the product of step (c);

.... .
. i,, .

(e) purifying the oxidized product of step (d) by RP-HPLC
using a bonded phase wide pore silica gel column and a solvent system containing acetonitrlle as an organic modifier and an organic acid selected from either heptafluorobutyric acid (HFBA) or trifluoroacetic acid (TFA);
(f) removing the organic solvents from the product of step (e~, recovering the IFN-B protein therefrom, and re-solubilizing the IFN-B protein in an appropriate buffer; and (g) further purlfying the product of step (f) by gel filtration.
Said process can further comprise the steps of formulating and optionally lyophilizing the product of step (9).
Preferably, the removal o~ the organic solvents and recovering the IFN-~ protein in step (f) is accomplished by 15 precip~tating the IFN-B protein and then resolubilizing the prote~n.
Preferably, such precipltation is accomplished by raising the pH and increasing the salt concentration. The salt concentration and pH are preferably raised by adding 3 M sodium phosphate~ pH 6, to the reverse-phase pool until a heavy precipitate formsO
Further preferred processes of the instant invention include those wherein ~he oxidizing step (d) of the above-described process is a controlled oxidation step carried out by using iodosobenzoic acid according to the method described in U.S. Patent No. 4,530,787 to Shaked et al., or by using an oxidation promoter containing a Cu+2 25 cation such as CuCl2 according to the method described in U.S. Patent NoO 4,572,798 to Koths et al. Preferably, the controlled oxidation step is carried out employing ~odosobenzoic acid.
The gel f~ltration of steps (c) and ~9) are preferably carried out on a S-200 column.

30 Process I
Another preferred purification process incorporating a RP-HPLC method option herein disclosed comprises, in addition ~o the process steps described immediately above, further process steps under 17 ~ 37 (a) for recovering the refractile material. Those steps are as follows:
6. solubilizing the refractile material under reducing conditions;
57.organically extracting the solubilized refractile material; and 8.isolating said refractile material ~rom the extractant.
Step 8 is preferably carried out by employing an acid precipitation step followed by centrifugation. The process containing such steps is herein designated Process I.

Process Il Another preferred alternatlve process of this invention for purifying recombinant IFN-~ includes the above-described steps respectively, for the process designated Process I except that steps (e) and (f), the RP-HPLC step and organic solvent removal stepS
respectively, occur before step (d), the controlled oxidation step.

Process III
Another preferred alternative process includes all the steps in the order of Process Il expect that step (c) (the first gel filtration step) is omitted.

Process IV
Ano~her preferred alternative process is similar to Process III except that the controlled oxidation step occurs before the RP-HPLC and removal of organic solvents.
- - .
Process V ~ -Another preferred alternative process is similar to Process : III except that an immunoaffinity purification step replaces the gel filtration step (9).

`

Process YI
Ano~her preferred alternative process of this invention is similar to the steps of Process II except that steps 6, 7, and 8 of Process II are omittedO
;

S Proces~ VII

I Another representative process of the present invention is similar to Process I except for the following differences: 1) there is an additinnal gel filtration step, preferably on an S-200 column, between the controlled oxidation step (d) and the RP-HPLC step (e);
10 and 2) the last gel filtration step (9) is preferably carried out on a G-75 column.

Process VIII
. _ _ Another preferred alternative process of this invention is similar to that of Process VII except that the third gel flltratton 15 step, preferably on a G-75 column, occurs prior to the RP-HPLC and removal of organic solvent steps rather than thereafter as in Process VII .

Preferred Yersion of Process I and, by Analogy, Processes II-VI
The steps of Process I and, by analogy, processes II-VI
20 could preferably be carried out as follows:
(a) growing the transformed bacterial hosts in an appropriate fermentation medium;
(b) concentrating the bacteria in the fermentat~on medium by cross-flow filtration, centrifugation or other conventional 25 methods;
(c)- disrupting the cell wall and cell membranP of the bacteria;
. (d) removing greater than 99% by weight of the salts from : said disruptate by diafiltration or centrifugation;

~e) redisrupting the desalted disruptate;

~95(:~37 (f) adding a ~aterial to the disruptate to increase the density or viscosity of, or to create a density or viscosity gradient in, ~he liquid within the disruptate;
(g) separating the refractile material containing the IFN-~5 protein from the cellular debris by high-speed centrifugation;
(h) solubilizing the refractile material in an aqueous buffer under reducing conditions;
(i) organically extracting the solubilized refractile materialJ preferably with 2-butanol or 2-methyl-2-butanol;
(j) isolating said refractile material from the extractant, preferably by employing an acid precipitation step followed by centrifugation;
(k) solub~lizing the resulting IFN-~ particle pellet with distilled water or with an aqueous solution of SDS at a IFN-~ to SDS
ratio of about 1:3;
(l) adjusting the pH of the solution to about 9O5 and reducing the solubillzed IFN-~;
(m) purifying the reduced IFN-B by gel chromatography;
~ n) collecting the eluted fraction of the purified IFN-~;
(o) oxidizing the eluate by a controlled oxidation method employing iodosobenzoic acid;
(p) purifying the oxidized product of step (o) by RP-HPLC
using a bonded-phase, wide-pore silica gel cotumn and a solvent system containing acetonitrile as an organic modifier and an organic acid selected form either HFBA or TFA;
.
(q) removing the organic solvents from the product of step (p) by precipitating the protein and thereafter recove~ing the protein by re-solubilizing it in an appropriate buffer;
. (r) further purifying the IFN-B by gel chromatography and collecting the eluate containing the purified IFN-~;
(s) desalting the purified IFN-B eluate in a desalting column e~uilibrated and run in 0.1% sodium laurate in 10mM Tris-HCl at pH 9.2;

(t) lowering the pH of the eluate rapidly to pH 3.0 with an appropriate acidic agentj (u~ centrifuging and filtering the IFN-~ pool;
(v) adding an effective amount of a non-ionic biodegradable polymeric detergent-con~aining solubilizer/stabilizer;
(w) adjusting the pH of the solution to abJut physiological pH;
(x) adding an appropr~ate polyol bulking/stab~l~zing agent in a concentration of from about 0.25% to about 10X;
(y) filtering the solution; and (z) immediately lyophilizing the IFN-B sample~
The lyophil k ed IFN-~ samp1e may be lyophllized if desired~
Ten m~l dithiothreitol (DTT) may be optionally included in the initial solubilizatlon step, and the mixture may be heated to about 50C for about 10 m~nutes. Preferably, 50 mM DTT can be included in said inital solubilization, and the mixture is heated to about 50C for about 20 minutes~
The IFN-~ is pre~erably oxidized in step (oj so that its cysteine residues are bridged to form cystines, as described by U.S.
Patent No. 4,530,787 to Shaked et al., using o-iodosobenzoic acid solution or by U.S. Patent No. 4,572,798 to Koths et al. using copper chloride. Preferably, o-iodosobenzoic acid is employed for such a controlled oxidation step.
European Publ~cation No. 843,997 publ~shed March 25, 1986 25 details a procedure for culturing a microorganism host transformed to produce IFN-~ and for extracting, purifying and formula~ing the ~ecombinant protein. Said purification procedure is exemplary of the processes in which the RP-HPLC method of the presen~ invention is ~ncorporated. A synopsis of said procedure follows.
The transformed microorganisms are grown in ~a suitable growth medium, typically to an optical density (OD) of at least about 30 at 680 nm, and preferably between about 20 and 40 at 680 nm. The composition of the growth medium will depend upon the particular ~ 37 microorganism involved. The medium is an aqueous nedium containing compounds that fulfill the nutritional requirements of the microorganism. Growth media will typically contain assimilable sources of carbon and nitrogen, energW sources~ magnesium, potassium and sodium ions, and optionally amino acids and purine and pyrimidine j bases. (See Review of Medical Biology, Lange Medical Publications, ~ 14th Ed pp. 80-~5 (1980).) In expression vectors involving the trp ¦ promoter~ the tryptophan concentration in the medium is carefully I controlled to become limiting at the time IFN-~ expression is desired. Growth media for E. coli are well known in the art.
After the cells are harvested from the culture, they may be concentrated, if necessary, to about 20 to 150 mg/ml, preferably 80 to 100 mg/ml (OD 40 to 300? preferably 160 to 200 at 680 nm), by cross-flow filtration, centrifugation, or other conventional methods.
Preferably, a compound which is non-toxic to humans, such as 1-octanol, in an amount of about 1% by weight of total components, is added to the fermenter before or during cell concentration to ensure that no v~able recombinant organisms remain before containment is broken.
Following concentration of the harvested culture, the cell membranes of the microorganisms are disrupted. Conventional cell disruption techniques such as homogenization, sonication, or pressure cycling may be used in this step of the process. Preferred methods are sonication or homogenization with a homogenizer. The end point of the disruption step can be determined by monitoring the optical density with the absorbance at 260 nm of the suspension typically increasing with cell lysis. In any event, the disruption should break substantially all of the cells so that substantially no intact cel7s are carried ~hrough to the solubilization step. Before the 30 disruption~ the pH of the liquid phase of the concentrate is adjusted, if necessary, to a 1evel that facilitates removal of E. coli proteins in subsequent steps3 while retaining the heterologous protein as an insoluble complex in the cellular debris.
After the cells have been disrupted, deionized water is 35 preferably added to the disruptate and greater than 99% by weight of ~he salts are removed therefrom. The removal of these salts to reduce the ionic strength of the disruptate may be accomplished by diafiltration using deionized water to flush out the ions or by centrifuging to pellet the cellular debris and refractile bodies followed by resuspension in deionized water.
After the salts are essentia11y removed, optionally a compound such as l-octanol may be added to the desalted disruptateJ if not added earlier, to ensure that no viable recombinant organisms remain. The desalted disruptate is again disrupted as described above for the initial disruption.
After redisruption, density or viscosity is increased and/or a gradient is created during centrifugation in the liquid within the disruptate by adding a material to the disruptate. One means to accomplish this goal is to add a material such as a sugar or mTxture 15 of sugars or a two-phase system, such as a glycerol/sugar mlxture which increases the density of the liquid to a p of about 1.1 to 1.3 g/ml, preferably 1.13 to 1.17 glmi. Also, the viscosity of the liquid phase may be increased to from 5 to 10 centipoise by any suitable means such as by adding a viscous compound such as, e.g., sucrose or 20 glycerol, thereto.
In the final step of the abbreviated "front-end" process to recover the refractile bodies, the refractile bodies containing the desired protein are separated from the cellular debris by hi~h-speed centrifugation. By "high-speed centrifugation" is meant spinning the 25 suspension in a centrifuge at about 10,000 to 40,000 times gravity, preferably about 10,000-20,000 x 9, for a suitable time period depending on the volume, generally about 10 minutes to seventy-two hours. The pellet resulting from the centrifugation is called the "particle pellet" or ~particle paste." The abbreviated front-end 30 process is most preferably used when sodium laurate is the primary solubilizing agentO
In an alternative, expanded "front-end" process to recover the refractile bodies, the particle pellet obtained from the last centrifugation step of the abbreviated front-end process is 35 solubilized, reduced and then extracted ~rom the aqueous medium with ~Z~ 7 2-butanol or 2-n~thyl-2-butanol. The extractant phase is then precipitated ~ith an acid and centrifuged to produce a "final pellet"
or "final paste," which is then further purified as indicated.
The alternative, expanded front-end process is distinguished from the abbreviated front-end process in th~t it comprises several additional steps as follows- solubilizing th~ refractile bodies under reducing conditions; organically extracting tne solubilized refractile material; and isolating said refracti~e ma~erial from the extractant. Essentially, the enhanced purity of the final pellet as opposed to the particle pellet lessens the purifying burden of downstream processing. There is an interdependence between the choice of the front-end process and later process purification steps to achieve the desired purity level for the final product. Once the choice of the particular front-end recovery of the refractile bodies has been made, one skilled in the art can pick and choose the alternative purifying steps outlined below to achieve the des~red purity level of the final product.
Whether the abbreviated or expanded front-end process is utilized to recover the refractile bodies containing the IFN-B~ the next step in purification is solubilizing either the particle or final pellet containing the refractile material. The following solubilizing agents can be used: sodium dodecyl sulfate (SDS), sodium laurate, urea, sodium dodecyl sulfonate, sodium decyl sulfate, sodium tetradecyl sulfate, sodium tridecyl sulfonate, sodium dodecyl N-sarcosinate, sodium tetradecyl N-sarcosinate, sodium dioctylsulfosuccinate, and guanidine hydrochloride. Preferred solubilizing agents are SDS, sodium laurate or guanidine hydrochloride.
The solubilizing agent is in an aqueous buffer, preferably phosphate buffered saline. The preferred percentage of the solubilizing agent is in the range of 1% to 5X (w/v). (Percentages herein reflect weight to volume ratios.) The preferred solubilizing solutions are phosphate buffered sal~ine ~ith 1-2% sodium laurate ~20mM
NaP04) at pH 9-10 and 1-5% SDS in 50 mM ~-mercaptoethanol. Sonication is preferably employed when sodium laurate is employed as the solubilizing agent to promote solubilization.

~ ~3 ~ 7 Reducing agents that can be employed during the solubilization step include: ~-mercaptoethanol (~-mer), glutathione, cysteine and dithiothreitol (DTT). DTT and B-mer are the most preferred reducing agents. It is preferred that reclucing agents be employed when either sodium laurate or guanidine hydrochloride is used j as the primary solubilizing agent.
The solubilization will typically be carried out at temperatures in the range of 20C to 25C with miXill9 to facilitate contact between the solid phase and the solubilizing medium.
Optionally, a reduction step may be carried out at this point. The pH, if necessary, may be adjusted to a range of 8.5 to 10, most preferably approximately 9. The suspension may be heated to 50 t 5C
for 5 to 15 minutes under nitrogen. The reaction mixture would then be cooled to approximately 25C.
The solubilizat~on is considered complete when the sample has sat 15 minutes or the solution turns translucent. Optionally at this point~ the insoluble material may be separated by centrifugation or filtration after completing the solubilization.
After the protein is solubilized, the resulting suspension 20 may optionally be centrifuged at 10,000-40,000 x 99 preferably 10,000 to 15,000 x 9, to obtain a pellet containing, inter alia, additional host (e.g., E. coli) proteins, notably including certain contaminants that ildVe molecular weights very close to that of the desired protein. The exact speed of centrifugation is not critical, as most 25 of the insoluble material will come out, even at low speeds. The pellet is discarded and the supernatant containing the desired protein is retained and processed to recover the desired protein.
If a reduction step was not carried out during the solubilization, the next step in the process would be a reduction of 30 the solubilized refractile body protein. A preferred reducing agent is dithiothreitol ~DTTj. Reduction conditions may also include the addition of a chelating agent such as ethylenediaminetetraacetic acid (EDTA).

~9~Q537 The next step in the process is to separate the protein in the supernatant from any host contaminants remaining after the centrifugation or filtration and optimally from the solubilizing agent. According to this invention, as represented by preferred embodiments of the instant invention outlined above, combinations of gel filtration, RP-HPLC, and/or immunoaffinity chromatography can be used. [Immunoaffinity chromatography is described in Pestka et al., Pharmac. _Ther~, 29:299-319 (1985).~ Gels that are capable of fractionating the solution to permit separation of the protein from these contaminants are commercially available. Sephacryl~ S-200 is a preferred gel for removing the higher molecular weight components, and Sephadex~ G-50, G-75 or G-100 gels are preferred for removing the low molecular weight contaminants. The gel filtrations will typically be run in buffered solutions (pH 5.5 to 7.0) containing about 0.1~ to 1.5% solubiliz~ng agent and about 0.5 to 10 mM reducing agent. lhe column will be sized to permit suitable resolution of the desired components.
RP-HPLC is capable of re~oving molecules from the solution that have molecular weights close to the protein and cannot, 20 therefore, be removed completely by gel filtration. In addition, contaminants such as bacterial endotoxin are also removed effectively by RP-HPLC. Therefore, RP-HPLC may also be used as a final purification step after gel filtration.
It is preferred in carrying out the process of this 25 invention that the last step of purification before stabilization of the formulation is a desalting step employing a transfer component, such as sodium laurate at a pH range of about 8.5 to about 10. The purity of the protein after the chromatography step(s) is at least about 95% and- higher~ and usually at least about 98%. This highly 3~ pure materia! contains less than about 2 ng endotoxin, usually less than about 0.01 ng endotoxin per 100,000 units protein bioactivity.
The formulation of the protein may be carried out as a separate operation using purified, selectively oxidized protein or in an operation that is integrated with the purification of the 35 selectively oxidized protein.

Conventional solid non-protein bulking/stabili~ing agents that are used in pharmaceutical tablet formulation ma~ be used as the carrier. These materials are water soluble, do not react with the IFN-B protein, and are themselves stable~ They are a1so preferably non-sensitive to water, that is, non-hygroscopic. Spec~fic examples of candidate carriers include jpolyols, starches and starch hydrolysates derived from wheat, corn, rice, and potatoes, as well as micro-crystalline celluloses.
The unit dosage amounts, that is, about 0.125 to 2 mg, preferably 0.25 to l mg of the recombinant beta-interferon, are dispensed into containers, the containers are capped with a slotted stopper, and the contents are lyophilized using conventional freeze-drying conditions and apparatus.
The lyophilized, sterile product conststs of a mixture of (1) recombinant beta-interferon; (2) carr~er9 preferably a polyol, and more preferably, dextrose or mannitol; (3) a stabilizer, preferably HSA, or a non-ionic biodegradable polymeric detergent, preferably Trycol LAL-12 or Plurafac C-17; and ~4) a small amount of buffer that will provide a physiological pH when the mixture is reconstituted.
The product may also contain a minor amount of a preservative to enhance chemical stability.
The lyophilized mixture may be reconstituted by injecting a conventional parenteral aqueous injection such as distilled water for injection, Ringer's solution injection, Hank's solution injection, dextrose injection, dextrose and salt injectlon, physiological saline injection, or the like, into the vial. The injection should be added against the side of the vial to avoid excess foaming. The amount of injection added to the vial will typically be in the range of 1 to 5 ml, preferably I to 2 ml.
-30 The reconstituted formulation prepared as descr~bed above is suitable for parenteral administration to humans or other mammals in therapeutically effective amounts (i.e., ~mounts which eliminate or reduce the patient's pathological conditlon) to provide therapy - thereto~ IFN-B therapy is appropriate for anti-cancer, anti-viral and 35 anti-psoriasis treatment.

~2~ 37 Recombinant IFN-~ is a "lipophilic protein9" a term used herein to refer to a protein which is not so1uble or not readily solublP in an aqueous medium under ambient conditions of room temperature and atmospheric pressure at a pH of between about 6.5 and 7.8. The term "recombinant protein" refers to a protein which is produced by recombinant DNA techniques wherein generally DNA is inserted into a suitable expression plasmid which is inserted into a host organism not native to the DNA which is transformed to produce ~ ~ the heterologous protein. The host may be any organism foreign to the 10 DNA, such as, e.g., bacteria, yeast, viruses, and mammals, among other types of cells. Preferably the host is microbial, and most preferably bacterial.
As used herein, the term "IFN-~" refers to ~-interFeron or ~-interferon-like polypeptides produced by recombinant DNA techniques 15 and whose amino acid sequence is the same as or similar or substantially homologous to the unglycosylated and/or glycosylated native ~-interferon.
The precise chemical structure of the IFN-~ protein will depend on a number of factors. As ioni2able amino and carboxyl groups 20 are present in the mo1ecule, a particular IFN-B protein may be obtained as an acidic or basic salt, or in neutral form. All such preparations which retain their activity when placed in suitable environmental conditions are included in the definition of IFN-~proteins herein. Further, the primary amino acid sequence of the 25 protein may be augmented by derivatization using sugar moieties (glycosylation) or by other supplementary mole~ules such as lipids, phosphate, acetyl groups and the like, more commonly by conjugation with saccharides. Certain aspects of such augmentation are accomplished through post-translational processing systems of th 30 producing host; other such modifications may be introduced in vitro.
In any event, such modifications are included in the definition of IFN-~ protein herein so long as the activity of the protein, as defined above~ is not destroyed. It is expected, of course, that such modifications may quantitatively or qualitatively affect the activity, 35 either by enhancing or diminishing the activity of the protein in the various assays.

~L~9~7 Fur~her, indiv~dual amino acid residues in the chain may be modif~ed by oxldat~on, reduction, or other derivatization, and the prote~n may be cleaved to obtain fragments which retain activity.
Such altera~ions which do not destroy activity do not remove ~he protein sequence from ~he definition.
jMost preferably the IFN-~ protein is unglycosylated IFN-B
which isjproduced by a microorganism that h~s been transformed with an IFN-~ gehe or a modification of the IFN-B gene that encodes a protein having: (a) an amino acid sequence that is at least substantially identical to the amino acid sequence of native IFN-B and (b~
biological activity that is common to native IFN-B. Substantial identity of amino acid sequences means the sequences are identical or differ by one or more amino acid alterations (deletions, additions, substitutions) that do not cause an adverse functional dissimilarity between the synthetic protein and the native IFN-~. Examples of such proteins are the IFN-~ prote~ns descrlbed in U.S. Patent Nos.
4~518,584 and 4,588,585. Most preferably, the IFN-~ is IFN-Bserl7 wherein ~he cysteine residue at amino acid position 17 is replaced by a serine residue.

EXAMPLES
The following examples further illustrate the RP-HPLC method and recombinant IFN-B purification processes of the invention. These examples are not intended to limit the invention in any manner. In these examples all temperatures are in degrees Celsius unless otherwise indicated-Representative Process for Purifying - Reco~binant IFN-B not Incorporating RP-HPLC
This example- provides a represen~ative process for recovering, purifying and formulating microbially produced IFN-~wherein a RP-HPLC method of the present invention is not employed.
This process employs the expanded front-end for recovering the refractile ma~erial containing the IFN-B protein and corresponds in outline to Protocol 3.

An analog IFN-~ designated IFN-~ser17 was recovered from E.
c _ . The amino acid sequence of this recombinant IFN-~ is different from that of native human IFN-~ in that the cysteine at position 17 has been changed to serine~ The strain of IFN-~serl7 producing E.
coli (K12/MM294-1) carrying plasmid pSY2501 used in this example was deposited at the American Type Culture Co11ection on November 18, 1983 under accession number 39,517. Said analog is described in U.S.
Patent Nos. 4,518,584 and 4,588,585 assigned to Ce~us Corporation.
The E. coli thus transformed were gro~n in a 1000-liter fermentor at 37C. The dissolved oxygen was maintained at about 40%
by, as necessary; (1) increasing agitation; (2) adding air; and (3) adding oxygen.
Once the fermenter was filled with water to the operating volume, the following trace elements were added:

ZnS04 7H20 72 mM
MnS04 . 4H20 30 ~M
CuS04 5H20 3 ~M
Na3 citrate 2H20 1.5 mM
KH2P04 21 mM
(NH4)2~4 72 mM

The fermenter feed and addition vessels were then sterilized according to standard operating procedures. Then the following sterile additions were made:

MgS04 7H20 20 mM
FeS04 . 7H20 100 ~M
L-tryptophan 70 mg/L
thiamine HC120 mg/L
glucose 5 gJL

The fermenter was cooled and inoculated with frozen or seed E. coli culture at 2 mg/L. A glucose feed was employed to maintain the glucose concentration between 5-10 g/L. At approximately 15 hours ~ 7 after fermentation was begun, the pH was adjusted wlth KOH to 6~8.
Op~ical density ~easurements and residual glucose measurements on samples were taken at 14-16 hours and approximate1y one hour intervals thereafter.
Induction of IFN-~ser17 production by depletion of L-tryptophan from the culture mediu~ occurred at about OD680=10 followed by the addition of casamino aci~s to a final concentration of 2~ at D680 15- The cultures werel harvested when glucose consumption reached 40 ~ 6 gfl.

Cell Concentration and Cell Wall and Membrane Disruption The refractile bodies containing the IFN-~ser17 protein were then isolated. The harvested mater~al was concentrated about 5-10 fold by circulating khe harvest material under pressure through UF
cross-flow filtra~on cartr~dges with a lOOK molecular we~ght 15 cutoff. Cells were d~srupted by 3 passes through a high-pressure homogen k er at 6,000 to 8,000 psig.

Diafiltration EDTA was added to the disruptate to a final concentration of mM. The suspension was then diafiltered against 5 volumes of 20 deionized water.

Redisruption EDTA was then added to a final concentration of 2 mM.
Octanol was added to lg (v/v) to kill any residual live bacteria in the diafiltered product. The suspension was redisrupted by passing it 25 twice through the hiyh-pressure homogenizer at 6,000-8,000 psig.

Sucrose Suspens~on and Centrifugation - Sucrose was added to the redisruptate to a findl concentration of 23% (wt/wt), creating a final density gradient between 1~1 and 1.25 g/ml. The mixture was centrifuged at 10,000 to 30 15,000 x 9, and the particle pellet or paste WdS collected. A

~295~37 temperature of at least 20C was maintained priGr to and during centrifugation.

Particle Paste Solubilizat~on and Reduction The par~icle pellet was then solubilized in phosphate buffered saline with 2X SDS. Solid DTT and EDTA were added to a final concentra~ion of 10 mM and 2 mM, respectively. The suspension was heated to 50 ~ 5C for 10 minutes under nitrogen~ The reaction mixture was then cooled to approximately 25C, and ~hen the pH of the mixture was adjusted to 7.4.

Or~anic Extraction and Acid Precipitation A volume of 2-butanol equal to the total volume of the suspension was measured. The suspension and organic solutiorl were pumped separately but simultaneously at flow rates of 1.1 to 1.3 liters per minute through a static mixer and then into a continuous centrifuge tat approximately 11~770 x 9) for phase separation. The 2-butanol-rich phase containing the IFN-~Serl7 was collected (Organic Extract).
The 2-butanol extract was mixed with 2.5 volumes of 0.1X SDS
in phosphate-buffered saline. Solid DTT was added to a ~inal concentration of 2 mM. lhe pH of the organic extract/buffer solutions was adjusted to 6.2 ~ 0.1 with glacial acetic acid (Acid Precipitate).

Centrifugation The mixture was then centrifuged (at 13,200 x 9) for approximately 2-6 hours, the supernatant was decanted and the final pellet was then collected (Final Pellet) containing approximately 81%
IFN-~. The final pellet containing the refractile material was then further purified by downstream process~ng.

Acid Precipitate Solubilization and Reduction ~RSAP) The final pellet was then re-suspended with 2~ SDS in 50 mM
30 phosphate buffer and 5 mM EDTA. Solid DTT was added to a final concentration of 20 mM, and the pH was adjusted to 8.5 with NaOH. The suspension was heated to 50 + 5C for 10 minutes under nitrogen, and then cooled to approximately 25C. The pH was then adjusted to a pH
of 5.5 with glacial acetic acid, and the solution was filtered through a 0.65 ~m filter. At this point the IFN-~ is in the form known as resuspended acid precipitate (RSAP).

S-200 Pre-Column The filtrate was then processed by pre-column chromatography by loading a Sephacryl~ S-200 pre-column and collecting fractions into clean, depyrogenated vessels using an elution buffer that is composed of 50 mM acetate, pH 5.5, 1 mM EDTA and 1% SDS. The fractions containing the IFN-~ monomer were pooled. I
The pre-column pool was then concentrated by using a hollow-fiber ultraflltration unit with a lOK molecular weight cut-off.

Oxidation Resulting in Oxidlzed Pre-column Pool The concentrated pre-column pool was then oxidized using o-iodosobenzoic acid (IBA). The oxidation was effected by adding equimolar amounts of protein and IBA into a reaction vessel containing 2 mM sodium pyrophosphate, 0.1~ SDS and 1 mM EDTA. A 20 ~M excess of IBA was present at the end of the oxidation. The pH was controlled at 9.0 + 0.1 with NaOH during oxidation, and adjusted to 5.5 + 0.2 with glacial acetic acid when the oxidation was completed.

Concentration The IFN-~ protein was then concentrated using a hollow-fiber ultrafiltrit~on unit with a lOK molecular weight cut-off.

S-200 Main Colùmn The protein was then loaded onto the main column (Sephacryl~
S-200), and fractions were collected into clean, depyrogenated vessels using an elution buffer that is composed o~ 50 mM ace~ate~ pH 5.5, 1 mM EDTA and 0.1% SDS.

A SDS-PAGE was performed on samples from each fraction tube starting from the beg~nniny of the peak to be pooled to the end of the peak. Using the SDS-PAGE results, the fractions containing no high molecular weight contaminants were determined Those fractions were then pooled.

Goncentration ,' The main ~olumn pool was then concentra~ed by using a hollow-fiber ultrafiltration unit with a lOK molecular weight cut-off.

G-75 Column The above procedure performed with the main column was repeated on a Sephadex~ G-75 column. Us1ng the SDS-PAGE results, the fract~ons containing neither low nor high molecu'lar weight contaminants were pooled.

G-25 Column and pH Ad~ustment The desalting step was then performed at p~ 9.2 wherein 0.1%
sodium laurate was used as a transfer component as follows. The pH
was adjusted wlth an appropriate basic agent such as lmM NaOH~
A Sephadex~ G-25 column was then equibrated with 0.1% sodium laurate in 10 mM Tris-HCl, pH 9.2 and loaded with the Sephadex~ G-75 20 pool containing 0.1~ SDS. Using the process chromatogram, t~e IFN-~serl7 peak was collected. The pH of the eluate was then lowered quickly with 1.0 N HC1 to pH 3.0, which precipitated the sodium laurate, but left the IFN-~Serl7 in solution.

Centrifu~ation and Filtration - The mixture was centrifuged at 35,000 x 9 for 30 minu~es and the supernatant was filtered through a 0.22 micron nitrocellulose f~lter. SDS concentration was assayed by acridine orange. ~Sokoloff et al., "Rapid -Spectrophotometric Assay of Dodecyl Sulfate Using Acridine Orange," Anal. Biochem., 118:138-141 (1981).3 The recovery 30 of the G-25 pool was above 85%, and the SDS concentration was reduced ~to less than 10 ~gjm~. ~

Formulat~on ~ _ ..
a The filtered supernatant was then stabili ed by adding 0.15%
.s Trycol LAL-12. The pH of the formulated product was then raised to about 7.0 ~ 0.3 with NaOH. The bulking/stabilizing agent, 5%
dextrose, was then added. Thc solution was then pre-filtered and sterile filtered through 0.45 and 0~22 micron nitrocellulose filters9 respectively. Immediately thereafter the correc~ dosage amounts of the IFN-Bser17 0.25 mg containing 0.5 x 108 units~ were aseptically filled into sterilized vi~ls with sterilized stoppers under sanltary and sterile conditions that were carefully monitored. The vials were then quickly placed in a lyophilizer where appropriate thermocouples were attached. The vials were frozen to between -35 and -45C. The lyophilization cycle was completed, and the vials were mechanically sealed under a vacuum.

Reducinq Levels of Microimpurlties by RP-HPLC
Approximately 50 mg of IFN-B protein eluted from the G-25 column (see E~ample 1) was concentrated in a stirred Amicon~ cell to 17.4 mg/ml. A 3 ml aliquot (52.2 mg) was applied to a 22 mm C18 column (15-20 micron particles) that had been equilibrated in 50g acetonitrile, 0.1~ HFBA. The IFN-~ protein was eluted with a linear gradient of a 0.33% increase in acetonitrile concentration per minute ard with a flow rate of 15 ml/min. Hal~-minute fractions were collected, beginning at 30 minutes. The eluted fractions were analyzed in two ways: by Western blotting and by reinjection into an analytical C18 column.
For Western blot analysis, fractions were -evaporated to dryness in the presence of a small amount of SDS, usiny a concentrator, and resuspended in 1% SDS. The Western blots were run and developed for IFN-~ (monoclonal anti-IFN-~; 0.5 ~g protein/lane) or E. coli (rabbit anti-E. coli protein; 10 ~g protein/lane).
_ The Western blots of the IFN-B species in the various fractions wherein the samples were reduced (UREA/SDS-PAGE) or non-~f~ k reduced (SDS-PAGE) indicated that the low molecular weight species were most concentrated in ~ractions from the trailing edge of the main peak (peak B); whereas oligomers were concentra~ed in fractions both from the leading and trailing edge of the main peak (peak B). The mid-peak fractions were relatively clean of oligomers.
Western blots were run of the E. coli species present i~n each fraction and the G-25 starting material. The amounts of E.
protein species were so low, even in the 6-25 material, that it ~as difficult to draw conclusions.
For analytical RP-HPLC analysis, aliquots containing approximately 20 ~g of protein were reinjected on an analytical C18 column (5 micron particle size). The samples were eluted using a 50-65% linear gradient of acetonitrile/0.1% HFBA.
Peaks A and C on the chromatographs represent m~nor IFN-~
lS species that are considered microimpurities whereas peak B containsthe main IFN-~ product. The amounts of peaks A and B in each fraction and in the G-25 starting material are listed in Table 1. The results indicate that the proportion of peak A decreases and the proportion of peak B increases progressively through the chromatogram.

Table 1 Peak A, Peak B, and Retention Time of the Major Peak in Each Fraction Peak A (X)Peak B (g) T of Major Peak Fraction ~ (2~.4-25.8 Minutes) ~Minutes) 6-25 (Concentrated 6.1 58.5 25.6 to 17.2 mg/ml) 32.5 Minutes 99.8 Below Detection Limit 20.0 36.0 Minutes 4.2 Below Detection Limit 24.5 38.0 Minutes 2.1 49.8 25.4 39.0 Minutes 0.3 69.4 ?5.4 40.0 Minutes ~.3 98.7 25.4 4?.0 Minutes 0.3 97.7 25.6 45.5 Minutes Belo~ Detection 100.0 25.8 Limit ~2g~7 This example indicates that the RP-HPLC method gives a good separation of many of the IFN-~ species seen by both analytical RP-HPLC and by Western blotting. Compared to the G-25 starting material, the mid-peak fractions obtained by this RP-HPLC separation were reduced in peaks A, C ard other species from the leading and trailing edges of peak B9 as ~ell as low molecular weight IFN-~ fragments and oligomersO 1l Purification of G-75 Pool IFN-~ by Pre~arative Scale RP-HPLC
This example indicates that when one-half gram of G-75 IFN-~
(see Example 1) was purified by a RP-HPLC method of this invention that the reverse phase pool contained approximately one-third as much peak A as the starting material and was reduced in peak C and other species detected by analytical RP-HPLC, as well as in the low molecular welght species and d~mers detected by Western blotting.
Whereas Example 2, where~n 50 mg of G-25 IFN-~ was eluted over a C18 column, represents an intermediate step in scaling up a RP-HPLC purif~cation of IFN-~ to the preparative level, in this example~
the RP-HPLC load was increased 10-fold to a load comparable to that which would be used in a preparative separation process. G-75 IFN-e prepared according to Example 1 was used as the starting material.
The experiment was performed by first concentrating G-75 IFN-B to 23~8 mg/ml in a stirred Amicon cell. Five aliquots of 5 ml each, totalling 476 mg of protein, were applied to a pre-packed C18 column (15-20 micron particles) that had been equillbrated tn 50%
acetonitrile/~.1% HFBA. The elution was in a gradient of from 50g to 80% acetonitrile with a 0.33~ increase in concentration of acetonitrile per minute and at a flow rate of 15 mltmin. Half-minute fractions were collected. Fractlons taken at 5-minute intervals were analyzed by Western blotting and analytical RP-HPLC às describe~ in Example 2.
To achieve a desired yield of 66% of the total amount of - IFN-~ protein eluted~ fractions from 55-80 minutes were pooled. The ~;Zffl~ 7 pooled fractlons were analyzed by Western blottlng and by analytical RP-HPLC using a computerized exponential skim integration method.
Recovery of protein as judged by A280 readings was approximately 100g. The column was washed by applying 5 ml of 1X SDS
and e1uting with the sa~e gradient. Only a very tiny peak of absorbance at 280 nm was eluted.
Table 2 summarizes a~d quantitates the results of the RP-HPLC analyses of the G-75 starting material, and of selected ~ractions.

Table 2 Analysis o~ Selected Fractions by RP-HPLC
Between Peak A (%) A & B (X) Peak 8 (%) "Post B" (%) Fract~on (18-20 (20.1-24.9 (25-26.5 (26.5+
15(Minutes) Minutes) Minutes) Minutes) M~nutes) G-75 IFN-~
DC-100 1.4 1.5 97.1 --Conc. G-75 IFN-~, DC-100 1.5 2.0 96.5 --32.5 3.3 76.3 14.7 5.8 37.5 2.9 96.5 -- 0.6 42.5 17.6 82.5 -- 0.2 47.5 3.6 19.2 77.0 0.1 52.5 0.5 2.8 96.7 --57.5 -- 1.2 98.4 0.4 62.5 -- 1.0 98.3 0.7 67.5 0.2 0.7 98.2 0.9 72.5 -- -- 100.0 --77.5 -- -- 100.0 --82.5 ~ 97.2 2.8 87.5 -- 27.9 72.1 --~2.5 3.9 37.0 51.6 7.5 96.5 ~ - 100.0 97.5 -- -- -- 100.0 ~2~ 7 As the results in Table 2 show~ there is a progressive decrease in the amount o~ peak A and a corresponding increase in the percentage of peak B in the fractions collected from 42.5-82.5 minutes, as well as a rise in the quantity of ~post B" species in the 5 82~5 minute fraction.
Western blots for IFN-B protein indicated t~at the low molecular weight IFN-~ species are concentrated in fract~ons from 37-47 minutes. Whereas said blots indicated that dimer~ and higher molecular weight IFN-B species are concentrated in both the earlier and later fractions, the fractions from the ~id-portion of the chromatogram are reduced in these species. No separation of the high molecular weight IFN-B band from the main band of IFN~B monomer was achieved.
A Western blot for E. coli fractions indicated that these proteins appear to be concentrated ln the early fractlons, while none were detected ~n fractions from 52 minutes and later. It was, however, difficult to draw conclusions, in that the levels of the E.
coli proteins were so low in the starting material.
Fractions from 55-80 minutes-were pooled to give a yield of 66% of the eluted protein. RP-HPLC analysis of the pooled fractions and starting material indicated that the poo~ed material fro~ the RP-HPLC column was reduced in peak A as compared to the G-75 IFN-~.
The amounts of peak A, peak B, and other species, as quantified by a computerized exponential skim integration, are listed 2~ ~n Table 3, below.

Table 3 Analysls of Pooled Fractions by RP-HPLC
Peak A (%) Peak B (X) (30.0-30.1(32.5-32.6Other Species ~ ~ Minutes) G-75, DC-100 ' (3~13 mg/ml) 2.8 92.1 5.1 I Conc. C-75, -. - j DC-100 (23.8 mg/ml) 9.2 82.8 8.0 RP Pool1/
(1.51 mglml) 0.9 95.2 3.9 1I Fractions from 55-80 minutes were pooled. Yield = 66.6~
The pooled fractions from the RP-HPLC elution contain approximately one-th~rd the amount of peak A, as does the G-75 IFN-~, and are re~uced in other minor species of IFN-~. Further, Weskern blots showed that the pooled fractions contain less of the low molecular weight species and dimers of IFN-~ than does the 6-75 IFN-~

PreParative Scale RP-HPLC Run of Oxidized Pr~ 5~ IFN-~
This example indicates that a preparative-scale RP-HPLC
separation of oxidized pre-column pool (see Example 3~ on a 22 mm C18 column, wherein the solvent system is acetonitrile/0.1~ HFBA in a linear gradient, results in separations of peak A and other species detected by analytical RP-HPLC and of low molecular weight fragments, dimers, and oligomers seen by Western blotting. The elution patterns of these species are similar to the elution patterns seen with G-75 - IFN-~ (see Example 3), which is a more pure starting material. ~ -- ~3~ The experiment was performed by first- ~iltering the oxidized pre-column IFN-~, prepared according to the procedures of Example 1, through a 0.45 micron filter. Then, 7S0 mg of the filtered, oxidized pre-column pool was pumped onto a pre-packed C18 column (22 mm diameter~ 15-20 micron particles) that had been equilibrated in 50g acetonitrile/0.1~ HFBA. IFN-~ was eluted with a biphasic solvent 35~`~37 gradient where1n sol~ent A was 0.1X HFBA in water and solvent B was 0.1% HFBA in acetonitrile 1n a gradient from 50% to 100%o The flow rate was 15 ml/minO Recovery of the IFN-~ protein, as measured by absorbance at 280 nm, was approximately 75~.
Fractions were collected at half-minute intervals. Specific activities of such fractions were determined by the cytopathic effect (CPE) assay according to Steward, W.E., The Interferon System (New York: Springer-Yerlag 1981) at page 17, and the results are listed below in Table 4 for fractions collected at five-minute intervals.

Table 4 IFN-~ Activity of Fractions from RP-HPLC of Oxidized ~re-Column Pool Fraction Spec~fic Activlty1/ (U/mq) Ox. Pre-Column Starting Mater~al 8.4 x 106 __ __ __ 8.0 x 104 7.0 x 106 5.0 x 107 1.4 x 108 100 2.0 x 108 110 1.8 x 108 120 1.7 x 107 13~ 2.4 x 108 140 3.3 x 108 150 3.3 x 108 1~ 3~5 x 108 170 - 6.5 x 107 180 7.7 x 107 1t CPE assay. All samples done in triplicateO

~2~5~7 Table 4 indicates that fractions from the middle of the chromatogram are approximately an order of magnitude more active than the starting material7 whereas the specific activities of the early and late fractions are less than or equal to that of the starting material. According to the CPE assay, approximately 700% o~ the original IFN-~ activity was reco~ered.
Selected fractions were analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and¦Western blotting.
For Western blot analysis, fractions' collected at five-minute intervals were evaporated to dryness in the presence of a smallamount of SDS, using a concentrator, and resuspended in 1% SDS.
Western blotting was performed with monoclonal anti-IFN-~ [(0.2 ~9 protein/lane); reduced samples (UREA/SDS-PAGE) and non-reduced samp'les (SDS-PAGE)] and with polycolonal anti-E. coli proteins (10 ~9 15 protein/lane). In the former blots for IFN-~ protein a high molecu'lar we~ght IFN-~ band was detected throughout the chromatogram, although it appeared more concentrated in the early fractions. The middle fractions contain the lowest amounts of impurities? lower and higher molecular weight species and dimers.
In the latter Western blot for E. coli proteins, the amounts of _ coli proteins were quite low in some of the middle fractions.
RP-HPLC ana~yses of fractions collected at five-minute intervals were per~ormed by reinjecting aliquots containing approximately 20 ~g of protein onto an analytical C18 column (5 micron 25 particle size). The samples were eluted using a 50-65X linear gradient of acetonitrile/O.lX HFBA. For comparison, G-75 IFN-~
prepared accordlng to Example 1 and the oxidized pre-column pool starting material for this example were also analyzed. The analytical RP-HPLC results are summarized quantitatively in Table 5.
.

12~5~ !37 Table 5 A~ounts of Peak A and Peak B in Fract~ons from the RP-HPLC of 750 mg Oxidized Pre-Column Pool on d 22 mm C18 Column Peak A (X)Peak B (%) Fraction (17.?-19.6 Minutes) (23.0-25.~? Minutes) G-75, DC 100 3.1 93.4 Ox. P.C.3 BC-203 2.1 8206 16.5 --41.3 9.6 _~
_ __ 0.4 92.0 105 0.~ 94.2 115 0.6 94.0 125 0.6 94.8 135 0.7 90.8 145 1.9 85.8 155 1.5 64.9 165 2.4 __ 175 14.8 --The analytical RP-HPLC and Western blots indicated that the preparative-scale RP-HPLC separation of oxidized pre-column .pool on a 22 mm C18 column results in separations of peak A and other species 25 detected by RP-HPLC and of low molecul.ar weight fragments, d~mers and oligomers seen by Western blotting... Yery early fractions are enriched.
: ~ ~ in peak A and species that elute between peak A and 8, whereas late fractions are enriched in post-peak B IFN-~ species. The elution patterns of these species are similar to the elution patterns with G-30 75 IFN-B~ a purer starting material. Further, the separation of the ~2~

high molecular weight IFN-~ band detected throughout the chromatograph for the RP--HPLC treated material was not seen when G-75 IFN-~ was used as the start~ng material, Further~ ~he Western blot for E. coli proteins Indicated that the prepar~ti~e scale RP-HPLC of the oxidized pre-column pool, representative of the improved RP-HPLC ~ethods disclosed herein, also separated E. coli proteins.

Purification of Resuspended Acid Prec1pitate ~RSAP) by Pre~arative Scale RP-HPLC and S-200 Gel Filtration This example demonstrates that an alternative purification process comprising the application of preparative amounts oF
resuspended ac~d precipitate (RSAP; see Example 1) on a 22 mm C18 column fullowed by oxidation of the reverse phase pool and further purification on a S-200 ma~n column, results in a product having a level of E. coli proteins comparable to that of the G-75 IFN-~ (see Example 1) and also a reduction in the proportion of peak A species.
Therefore, such a combination of RP-HPLC and S-200 gel filtration results in a more efficient process than that detailed in Example 1 and a purer product. This example also shows that the RP-HPLC method can handle a full preparative equivalent load of RSAP on a C18 column, provided the size of the load and the slope of the elution gradient are balanced.
The experiment was performed by applying 800 mg of the RSAP, prepared according to Example 1, to a 22 mm C18 column equilibrated in 50% acetonitrile/O.Ig HFBA startlng buffer.
The column was washed with starting solvent and then eluted with a linear gradient from 50-65X acetonitrile/b.1X HFBA~ -Because of 30 the ~size of the load~ the slope of the elution gradient was greatly reduced as compared to runs wherein les`s material was loaded on the reverse phase column. The slope of the elution gradient, instead of being, for example, a 0.1% increase per minute for a 580 mg load of RSAP, was, in this experiment, a 0.04% increase in acetoni~rile 35 concentration per minute~

~2~ 87 Fractions from the reverse phase column were evaluated by silver stained SDS Phast gels and by Western blotting for Eo coli proteins. The levels of E. co!i proteins at this point were shown to be comparable to levels wherefn the load of RSAP was smaller but the slope of the elution gradient was steeper as indicated above.
The run was completed by fu~rther purification on a S-200 main column. Protein from the revers;é phase pool was precipitated, resolubilized at a concentration of d.25 mg/ml~ and oxidized by the CuC12 method descr~bed in U.S. Patent No. 4,569,790. The oxidized material was then eluted on an S-200 main column.
The S-200 column fractions were examined by silver stained SDS Phast gels and pooled to eliminate those fractions that were highest in the high molecular we~ght IFN-~ monomer. Pools from each stage of the process were examtned by Western blotting for IFN-~ and lS E. col~ proteins and by analyt~cal RP-HPLC.
A comparison by analytical RP-HPLC of the S-200 pool from the test process w~th a sample of ~-75 IFN-B prepared according to Example 1 indicates that the material from the test process is clearly lower in peak A. The overall purity o~ the test process material, reported in Table 6, is increased from 92g to 99g peak B, according to a ~angential skim analysis from the analytical RP-HPLC chromatograms.

Table 6 Levels of Peak A and Peak B
~ Peak A (~) Peak B (~) S-200 Pool, Test Process 0.9 99.1 G-75 4.2 92.1 Western blots of both reduced and non-reduced samples from the test process for IFN-B prote~ns ~nd~cated that the S-200 pool is lower in dimers than the G-75 material produced according to Example 1.

The Western blots for E~ coli proteins, both from the S-200 poo1 of the test process and from the G-75 pool of Examp1e I, were both so low in E. coli proteins as to make comparisons difficult.
Thus, the test process was confirmed tn producing a product with very low amounts of E~ coli proteins. A conclusion from this experiment was that balancing the amount of material loaded on the reverse phase column with the slope of the elution gradient purification processes of this inven~ion results in a product which has therapeutically acceptable levels of E. coli proteins.

Conclusion In summary, it can be seen that the RP-HPLC methods herein described provide a new and effective means of purifying recombinantly produced IFN-~. Benefits o~ such RP-HPLC purification method include the removal of non-IFN-~ protein and bacterial endotoxins and the 15 reduction ln levels of m~nor IFN-~ species. Furt~er, puriflcation processes incorporating a RP-HPLC method option of this 1nvention either as an addition to other purification steps or as an alternative to one or more of such purification steps can result in very effective and efficient processes for purifying recombinant IFN-~.
The benefits of the alternative downstream purification processes herein described for IFN-~ include simplification of the downstream purification process; reduction in downstream process ti~e;
good reliability in terms of bacterial endotoxin removal; a process that can be reproducibly scaled-up; and a very pure IFN-B product wlth 25 reduced levels of minor IFN-B species.

Deposits As mentioned above, a culture of E. coli KI2/MM294-1 carrying plasmid pSY2501 was deposi~ed at the American Type Culture Collection 12301 Parklawn Drive, Rockville, MD-208529 US, on November 30 18, 1983 under ATCC No. ~9~517.
Said deposit was made pursuant to a contract between the ATCC and the assignee of this patent application, Cetus Corporation.
The contract with the ATCC provides for permanent availability of said strain and progeny thereof to the public upon issuance of a U.S.
patent related to this app1icat~on describ~ng and identify~ng the deposit or upon the publ~cation or laying open to the public of any . U.S. or foreign patent application, whichever comes first, and for the availability of the straln and the progeny thereof to one determined by the U.S. Commissioner of Patents and Trademarks to be entitled thereto according to 35 USC ~122 and the Commissioner's rules pursuant thereto (including 37 CFR 1.14 with particular reference to 886 OG
638). The assignee of the present application has agreed that if the strain on deposit should die or be lost or destroyed when cultivated under suitable conditions, it will be promptly replaced upon notification with a viable culture of the sam2 strain.
The deposit agreement under the terms of the Budapest Treaty assures that said culture deposited will be maintained in a viable and uncontaminated condition for a period of at least five years after the most recent request for the furnishing of a sample of the depos~ted microorganism was received by the ATCC and, in any case, for a per~od of at least 30 years after the date of the deposlt.
Availability of the deposited strain is not to be construed as a license to pract~ce the lnvention in contravention of the rlghts granted under the authority of any govern~ent in accordance with its patent laws.

Claims (37)

1. A method for purifying recombinant interferon-.beta. (IFN-.beta.) which comprises isolating the IFN-.beta. from hosts transformed to produce it and passing the isolated IFN-.beta. through at least one bonded-phase, wide-pore, silica gel, reverse-phase high performance liquid chromatography column using a solvent system comprising an organic modifier comprising from about 50% to 100%
heptafluorobutyric acid (HFBA) and trifluoracetic acid (TFA), wherein said solvent system is a gradient of acetonitrile.

2. A method according to claim 1 wherein the column is an alkane reverse phase column.

3. A method according to claim 2 wherein said alkane column is a C4, C8 or C18 column.

4. A method according to claim 1 wherein the concentration (v/v) of HFBA or TFA is from about 0.001% to about 2%.

5. A method according to claim 4 wherein the concentration (v/v) of HFBA or TFA is from about 0.05% to about 1%.

6. A method according to claim ; wherein the concentration of acetonitrile is from about 50% to about 65%.

7. A method according to claim 6 wherein the concentration (v/v) of HFBA or TFA is from about 0.1% to about 0.2%.

8. A method according to claim 1 wherein the gradient is linear and the slope of said gradient is balanced with the amount of material containing recombinant IFN-.beta. loaded on the reverse phase column.

9. A method according to claim 4 wherein the column is C18 and the acid is HFBA.

10. A method according to claim 7 wherein the column is C18 and the acid is HFBA.

11. A method according to claim 4 wherein the column is C4 and the acid is TFA.

12. A method according to claim 7 wherein the column is C4 and the acid is TFA.

13. The method of claim 1 wherein the recombinant IFN-.beta. is produced in a microbe and the isolation steps comprise (a) disrupting the cell wall and cell membrane of a host microorganism cell culture transformed to produce said IFN-.beta.;
(b) removing greater than 99% by weight of the salts from said disruptate;
(c) redisrupting the desalted disruptate;
(d) adding a material to the disruptate to increase the density or viscosity of, or to create a density or viscosity gradient in, the liquid within the disruptate; and (e) recovering refractile material containing the IFN-.beta. by high-speed centrifugation.

14. The method of claim 1 wherein the purification is carried out by (a) solubilizing the IFN-.beta. with an aqueous solution of a solubilizing agent which forms a water-soluble complex with the recombinant IFN-.beta., said solution containing a reducing agent;
(b) oxidizing the product of step (a);
(c) passing the oxidized product of step (b) through a bonded-phase wide-pore, silica gel, reverse-phase high performance liquid chromatography column using a solvent system comprising from about 50% to 100% acetonitrile and an organic acid selected from heptafluorobutyric acid (HFBA) and trifluoroacetic acid (TFA);
(d) removing the solvent system from the product of step (c), recovering the IFN-.beta. protein therefrom, and resolubllizing the IFN-.beta. protein in a buffer; and (e) further purifying the product of step (d).

15. A process for recovering and purifying microbially produced interferon-beta (I .beta. ) comprising:
(a) disrupting the cell wall and cell membrane of a host microorganism cell culture transformed to produce said IFN-.beta.;
(b) removing greater than 99% by weight of the salts from said disruptate:
(c) redisrupting the desalted disruptate:
(d) adding a material to the disruptate to increase the density or viscosity of, or to create a density or viscosity gradient in, the liquid within the disruptate;
(e) recovering refractile material containing the IFN-.beta. by high-speed centrifugation:
(f) solubilizing the IFN-B in the refractile material with an aqueous solution of a solubilizing agent which forms a water-soluble complex with the recombinant IFN-.beta., said solution containing a reducing agent;
(g) oxidizing the product of: step (f):
(h) passing the oxidized product of step (g) through a bonded-phase wide-pore, silica gel, reverse-phase high performance liquid chromatography column using a solvent system comprising from about 50% to 100% acetonitrile and an organic acid selected from heptafluorobutyric acid (HFBA) and trifluoroacetic acid (TFA):

(i) removing the solvent system from the product of step (h), recovering the IFN-.beta. protein therefrom, and resolubilizing the IFN-.beta. protein in a [anappropriate] buffer; and (j) further purifying the product of step (i).

16. The process of claim 15 wherein the IFN-.beta. is recombinant IFN-.beta..

17. A process according to claim 15 wherein the oxidizing step (g) is a controlled oxidation step carried out by using iodosobenzoic acid or by using an oxidation promoter containing a Cu2+ cation.

18. A process according to claim17: wherein the oxidation promoter containing a Cu2+ cation is CuC12 and after step (f) and before step (g) the IFN-.beta. is separated from the resulting solution in the presence of said reducing agent.

19. A process according to claim 18 wherein a gel filtration step is performed on a S-200 column in step (j) or between steps (g) and (h).

20. A process according to claim 17 further comprising after step (e) and before step (f) the steps of:
(e') solubilizing the refractile material under reducing conditions;
(e'') extracting the solubilized refractile material with an organic solvent;
and (e''') isolating said refractile material from the extractant.

21. A process according to claim 20 wherein step (e''') is carried out by employing an acid precipitation step followed by centrifugation.

22. A process according to claim 20 wherein the solvent system is removed in step (i) by precipitating the IFN-.beta. protein.

23. A process according to claim 22 wherein the IFN-.beta. protein is precipitated by raising the pH and increasing the salt concentration of the product of step (h) until a heavy precipitate forms.

24. A process according to claim 23 wherein the salt concentration is increased by adding 3M sodium phosphate at pH 6.

25. A process according to claim 15 wherein the bonded phase, wide-pore, silica gel column is an alkane reverse phase column.

26. A process according to claim 25 wherein said alkane column is a C4, C8 or C18 column.

27. A process according to claim 26 wherein the concentration (v/v) of HFBA or TFA is from about 0.001% to about 2%.

28. A process according to claim 27 wherein the concentration of acetonitrile is from about 50% to about 65% and wherein the concentration of (v/v) of HFBA
pr TFA is from about 0.1% to about 0.2%.

29. A process according to claim 28 wherein the column is C18 and the acid is HFBA.

30. A process according to claim 26 wherein the column is C4 and the acid is TFA.

31. A process according to claim 20 wherein steps (e) and (f) occur before step (d).

32. A process according to claim 18 wherein the step after step (f) and before step (g) is carried out by gel filtration and step (j) is carried out by gel filtration or by an immuno-affinity purification step.

33. A process according to claim 20 further comprising a gel filtration step between step (g) and step (h).

34. A process according to claim 32 wherein the gel filtration after step (f) and before step (g) and the gel filtration between steps (g) and (h) are carried out on 5-200 columns, whereas step (g) is carried out by gel filtration on a G-75 column.

35. A process according to claim34 wherein step (j) is carried out prior to step (h) and (i).

36. A process according to claim 15 further comprising the steps of stabilizing, and formulating the purified IFN-.beta. protein.

37. A process according to claim 36 further comprising the step of lyophilizing the formulated IFN-.beta. protein.